[proxmark3-svn] / armsrc / optimized_cipher.c

/*****************************************************************************
 * WARNING
 *
 * THIS CODE IS CREATED FOR EXPERIMENTATION AND EDUCATIONAL USE ONLY.
 *
 * USAGE OF THIS CODE IN OTHER WAYS MAY INFRINGE UPON THE INTELLECTUAL
 * PROPERTY OF OTHER PARTIES, SUCH AS INSIDE SECURE AND HID GLOBAL,
 * AND MAY EXPOSE YOU TO AN INFRINGEMENT ACTION FROM THOSE PARTIES.
 *
 * THIS CODE SHOULD NEVER BE USED TO INFRINGE PATENTS OR INTELLECTUAL PROPERTY RIGHTS.
 *
 *****************************************************************************
 *
 * This file is part of loclass. It is a reconstructon of the cipher engine
 * used in iClass, and RFID techology.
 *
 * The implementation is based on the work performed by
 * Flavio D. Garcia, Gerhard de Koning Gans, Roel Verdult and
 * Milosch Meriac in the paper "Dismantling IClass".
 *
 * Copyright (C) 2014 Martin Holst Swende
 *
 * This is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as published
 * by the Free Software Foundation, or, at your option, any later version.
 *
 * This file is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with loclass.  If not, see <http://www.gnu.org/licenses/>.
 *
 *
 *
 ****************************************************************************/

/**

  This file contains an optimized version of the MAC-calculation algorithm. Some measurements on
  a std laptop showed it runs in about 1/3 of the time:

	Std: 0.428962
	Opt: 0.151609

  Additionally, it is self-reliant, not requiring e.g. bitstreams from the cipherutils, thus can
  be easily dropped into a code base.

  The optimizations have been performed in the following steps:
  * Parameters passed by reference instead of by value.
  * Iteration instead of recursion, un-nesting recursive loops into for-loops.
  * Handling of bytes instead of individual bits, for less shuffling and masking
  * Less creation of "objects", structs, and instead reuse of alloc:ed memory
  * Inlining some functions via #define:s

  As a consequence, this implementation is less generic. Also, I haven't bothered documenting this.
  For a thorough documentation, check out the MAC-calculation within cipher.c instead.

  -- MHS 2015
**/

#include "optimized_cipher.h"
#include <stddef.h>
#include <stdbool.h>
#include <stdint.h>


#define opt_T(s) (0x1 & ((s->t >> 15) ^ (s->t >> 14)^ (s->t >> 10)^ (s->t >> 8)^ (s->t >> 5)^ (s->t >> 4)^ (s->t >> 1)^ s->t))

#define opt_B(s) (((s->b >> 6) ^ (s->b >> 5) ^ (s->b >> 4) ^ (s->b)) & 0x1)

#define opt__select(x,y,r)  (4 & (((r & (r << 2)) >> 5) ^ ((r & ~(r << 2)) >> 4) ^ ( (r | r << 2) >> 3)))\
	|(2 & (((r | r << 2) >> 6) ^ ( (r | r << 2) >> 1) ^ (r >> 5) ^ r ^ ((x^y) << 1)))\
	|(1 & (((r & ~(r << 2)) >> 4) ^ ((r & (r << 2)) >> 3) ^ r ^ x))

/*
 * Some background on the expression above can be found here...
uint8_t xopt__select(bool x, bool y, uint8_t r)
{
	uint8_t r_ls2 = r << 2;
	uint8_t r_and_ls2 = r & r_ls2;
	uint8_t r_or_ls2  = r | r_ls2;

	//r:      r0 r1 r2 r3 r4 r5 r6 r7
	//r_ls2:  r2 r3 r4 r5 r6 r7  0  0
	//                       z0
	//                          z1

//  uint8_t z0 = (r0 & r2) ^ (r1 & ~r3) ^ (r2 | r4); // <-- original
	uint8_t z0 = (r_and_ls2 >> 5) ^ ((r & ~r_ls2) >> 4) ^ ( r_or_ls2 >> 3);

//  uint8_t z1 = (r0 | r2) ^ ( r5 | r7) ^ r1 ^ r6 ^ x ^ y;  // <-- original
	uint8_t z1 = (r_or_ls2 >> 6) ^ ( r_or_ls2 >> 1) ^ (r >> 5) ^ r ^ ((x^y) << 1);

//  uint8_t z2 = (r3 & ~r5) ^ (r4 & r6 ) ^ r7 ^ x;  // <-- original
	uint8_t z2 = ((r & ~r_ls2) >> 4) ^ (r_and_ls2 >> 3) ^ r ^ x;

	return (z0 & 4) | (z1 & 2) | (z2 & 1);
}
*/

void opt_successor(const uint8_t *k, State *s, bool y, State *successor) {
	uint8_t Tt = 1 & opt_T(s);

	successor->t = (s->t >> 1);
	successor->t |= (Tt ^ (s->r >> 7 & 0x1) ^ (s->r >> 3 & 0x1)) << 15;

	successor->b = s->b >> 1;
	successor->b |= (opt_B(s) ^ (s->r & 0x1)) << 7;

	successor->r = (k[opt__select(Tt, y, s->r)] ^ successor->b) + s->l ;
	successor->l = successor->r + s->r;
}

void opt_suc(const uint8_t *k, State *s, uint8_t *in, uint8_t length, bool add32Zeroes) {
	State x2;
	for (int i = 0; i < length; i++) {
		uint8_t head;
		head = 1 & (in[i] >> 7);
		opt_successor(k, s, head, &x2);

		head = 1 & (in[i] >> 6);
		opt_successor(k, &x2, head, s);

		head = 1 & (in[i] >> 5);
		opt_successor(k, s, head, &x2);

		head = 1 & (in[i] >> 4);
		opt_successor(k, &x2, head, s);

		head = 1 & (in[i] >> 3);
		opt_successor(k, s, head, &x2);

		head = 1 & (in[i] >> 2);
		opt_successor(k, &x2, head, s);

		head = 1 & (in[i] >> 1);
		opt_successor(k, s, head, &x2);

		head = 1 & in[i];
		opt_successor(k, &x2, head, s);
	}
	//For tag MAC, an additional 32 zeroes
	if (add32Zeroes) {
		for(int i = 0; i < 16; i++) {
			opt_successor(k, s, 0, &x2);
			opt_successor(k, &x2, 0, s);
		}
	}
}

void opt_output(const uint8_t *k, State *s,  uint8_t *buffer) {
	State temp = {0, 0, 0, 0};
	for (uint8_t times = 0; times < 4; times++) {
		uint8_t bout = 0;
		bout |= (s->r & 0x4) << 5;
		opt_successor(k, s, 0, &temp);
		bout |= (temp.r & 0x4) << 4;
		opt_successor(k, &temp, 0, s);
		bout |= (s->r & 0x4) << 3;
		opt_successor(k, s, 0, &temp);
		bout |= (temp.r & 0x4) << 2;
		opt_successor(k, &temp, 0, s);
		bout |= (s->r & 0x4) << 1;
		opt_successor(k, s, 0, &temp);
		bout |= (temp.r & 0x4) ;
		opt_successor(k, &temp, 0, s);
		bout |= (s->r & 0x4) >> 1;
		opt_successor(k, s, 0, &temp);
		bout |= (temp.r & 0x4) >> 2;
		opt_successor(k, &temp, 0, s);
		buffer[times] = bout;
	}
}

void opt_MAC(uint8_t *k, uint8_t *input, uint8_t *out) {
	State _init  =  {
		((k[0] ^ 0x4c) + 0xEC) & 0xFF,// l
		((k[0] ^ 0x4c) + 0x21) & 0xFF,// r
		0x4c, // b
		0xE012 // t
	};

	opt_suc(k, &_init, input, 12, false);
	//printf("\noutp ");
	opt_output(k, &_init, out);
}

uint8_t rev_byte(uint8_t b) {
	b = (b & 0xF0) >> 4 | (b & 0x0F) << 4;
	b = (b & 0xCC) >> 2 | (b & 0x33) << 2;
	b = (b & 0xAA) >> 1 | (b & 0x55) << 1;
	return b;
}

void opt_reverse_arraybytecpy(uint8_t *dest, uint8_t *src, size_t len) {
	for (size_t i = 0; i < len; i++) {
		dest[i] = rev_byte(src[i]);
	}
}

void opt_doReaderMAC(uint8_t *cc_nr_p, uint8_t *div_key_p, uint8_t mac[4]) {
	static uint8_t cc_nr[12];
	opt_reverse_arraybytecpy(cc_nr, cc_nr_p, 12);
	uint8_t dest[] = {0, 0, 0, 0, 0, 0, 0, 0};
	opt_MAC(div_key_p, cc_nr, dest);
	//The output MAC must also be reversed
	opt_reverse_arraybytecpy(mac, dest, 4);
	return;
}

void opt_doTagMAC(uint8_t *cc_p, const uint8_t *div_key_p, uint8_t mac[4]) {
	static uint8_t cc_nr[8+4+4];
	opt_reverse_arraybytecpy(cc_nr, cc_p, 12);
	State _init = {
		((div_key_p[0] ^ 0x4c) + 0xEC) & 0xFF,// l
		((div_key_p[0] ^ 0x4c) + 0x21) & 0xFF,// r
		0x4c, // b
		0xE012 // t
	};
	opt_suc(div_key_p, &_init,cc_nr, 12, true);
	uint8_t dest[] = {0, 0, 0, 0};
	opt_output(div_key_p, &_init, dest);
	//The output MAC must also be reversed
	opt_reverse_arraybytecpy(mac, dest, 4);
	return;
}

/**
 * The tag MAC can be divided (both can, but no point in dividing the reader mac) into
 * two functions, since the first 8 bytes are known, we can pre-calculate the state
 * reached after feeding CC to the cipher.
 * @param cc_p
 * @param div_key_p
 * @return the cipher state
 */
State opt_doTagMAC_1(uint8_t *cc_p, const uint8_t *div_key_p) {
	static uint8_t cc_nr[8];
	opt_reverse_arraybytecpy(cc_nr, cc_p, 8);
	State _init = {
		((div_key_p[0] ^ 0x4c) + 0xEC) & 0xFF,// l
		((div_key_p[0] ^ 0x4c) + 0x21) & 0xFF,// r
		0x4c, // b
		0xE012 // t
	};
	opt_suc(div_key_p, &_init, cc_nr, 8, false);
	return _init;
}

/**
 * The second part of the tag MAC calculation, since the CC is already calculated into the state,
 * this function is fed only the NR, and internally feeds the remaining 32 0-bits to generate the tag
 * MAC response.
 * @param _init - precalculated cipher state
 * @param nr - the reader challenge
 * @param mac - where to store the MAC
 * @param div_key_p - the key to use
 */
void opt_doTagMAC_2(State _init, uint8_t *nr, uint8_t mac[4], const uint8_t *div_key_p) {
	static uint8_t _nr[4];
	opt_reverse_arraybytecpy(_nr, nr, 4);
	opt_suc(div_key_p, &_init, _nr, 4, true);
	//opt_suc(div_key_p, &_init,nr, 4, false);
	uint8_t dest[] = {0, 0, 0, 0};
	opt_output(div_key_p, &_init, dest);
	//The output MAC must also be reversed
	opt_reverse_arraybytecpy(mac, dest, 4);
	return;
}
Commit	Line	Data
c99dc845 MHS	1	/*****************************************************************************
	2	* WARNING
	3	*
17505ce2	4	* THIS CODE IS CREATED FOR EXPERIMENTATION AND EDUCATIONAL USE ONLY.
	5	*
	6	* USAGE OF THIS CODE IN OTHER WAYS MAY INFRINGE UPON THE INTELLECTUAL
	7	* PROPERTY OF OTHER PARTIES, SUCH AS INSIDE SECURE AND HID GLOBAL,
	8	* AND MAY EXPOSE YOU TO AN INFRINGEMENT ACTION FROM THOSE PARTIES.
	9	*
	10	* THIS CODE SHOULD NEVER BE USED TO INFRINGE PATENTS OR INTELLECTUAL PROPERTY RIGHTS.
c99dc845 MHS	11	*
	12	*****************************************************************************
	13	*
	14	* This file is part of loclass. It is a reconstructon of the cipher engine
	15	* used in iClass, and RFID techology.
	16	*
	17	* The implementation is based on the work performed by
	18	* Flavio D. Garcia, Gerhard de Koning Gans, Roel Verdult and
	19	* Milosch Meriac in the paper "Dismantling IClass".
	20	*
	21	* Copyright (C) 2014 Martin Holst Swende
	22	*
	23	* This is free software: you can redistribute it and/or modify
	24	* it under the terms of the GNU General Public License version 2 as published
298e1a2d	25	* by the Free Software Foundation, or, at your option, any later version.
c99dc845 MHS	26	*
	27	* This file is distributed in the hope that it will be useful,
	28	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	29	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	30	* GNU General Public License for more details.
	31	*
	32	* You should have received a copy of the GNU General Public License
	33	* along with loclass. If not, see <http://www.gnu.org/licenses/>.
17505ce2	34	*
	35	*
	36	*
c99dc845 MHS	37	****************************************************************************/
	38
	39	/**
	40
	41	This file contains an optimized version of the MAC-calculation algorithm. Some measurements on
	42	a std laptop showed it runs in about 1/3 of the time:
	43
	44	Std: 0.428962
	45	Opt: 0.151609
	46
	47	Additionally, it is self-reliant, not requiring e.g. bitstreams from the cipherutils, thus can
	48	be easily dropped into a code base.
	49
	50	The optimizations have been performed in the following steps:
	51	* Parameters passed by reference instead of by value.
	52	* Iteration instead of recursion, un-nesting recursive loops into for-loops.
	53	* Handling of bytes instead of individual bits, for less shuffling and masking
	54	* Less creation of "objects", structs, and instead reuse of alloc:ed memory
	55	* Inlining some functions via #define:s
	56
	57	As a consequence, this implementation is less generic. Also, I haven't bothered documenting this.
	58	For a thorough documentation, check out the MAC-calculation within cipher.c instead.
	59
	60	-- MHS 2015
	61	**/
	62
	63	#include "optimized_cipher.h"
b8e461ff	64	#include <stddef.h>
c99dc845 MHS	65	#include <stdbool.h>
c99dc845 MHS	66	#include <stdint.h>
c99dc845	67
c99dc845 MHS	68
	69	#define opt_T(s) (0x1 & ((s->t >> 15) ^ (s->t >> 14)^ (s->t >> 10)^ (s->t >> 8)^ (s->t >> 5)^ (s->t >> 4)^ (s->t >> 1)^ s->t))
	70
	71	#define opt_B(s) (((s->b >> 6) ^ (s->b >> 5) ^ (s->b >> 4) ^ (s->b)) & 0x1)
	72
	73	#define opt__select(x,y,r) (4 & (((r & (r << 2)) >> 5) ^ ((r & ~(r << 2)) >> 4) ^ ( (r \| r << 2) >> 3)))\
	74	\|(2 & (((r \| r << 2) >> 6) ^ ( (r \| r << 2) >> 1) ^ (r >> 5) ^ r ^ ((x^y) << 1)))\
	75	\|(1 & (((r & ~(r << 2)) >> 4) ^ ((r & (r << 2)) >> 3) ^ r ^ x))
	76
	77	/*
	78	* Some background on the expression above can be found here...
	79	uint8_t xopt__select(bool x, bool y, uint8_t r)
	80	{
	81	uint8_t r_ls2 = r << 2;
	82	uint8_t r_and_ls2 = r & r_ls2;
	83	uint8_t r_or_ls2 = r \| r_ls2;
	84
	85	//r: r0 r1 r2 r3 r4 r5 r6 r7
	86	//r_ls2: r2 r3 r4 r5 r6 r7 0 0
	87	// z0
	88	// z1
	89
17505ce2	90	// uint8_t z0 = (r0 & r2) ^ (r1 & ~r3) ^ (r2 \| r4); // <-- original
c99dc845 MHS	91	uint8_t z0 = (r_and_ls2 >> 5) ^ ((r & ~r_ls2) >> 4) ^ ( r_or_ls2 >> 3);
c99dc845 MHS	92
17505ce2	93	// uint8_t z1 = (r0 \| r2) ^ ( r5 \| r7) ^ r1 ^ r6 ^ x ^ y; // <-- original
c99dc845 MHS	94	uint8_t z1 = (r_or_ls2 >> 6) ^ ( r_or_ls2 >> 1) ^ (r >> 5) ^ r ^ ((x^y) << 1);
c99dc845 MHS	95
17505ce2	96	// uint8_t z2 = (r3 & ~r5) ^ (r4 & r6 ) ^ r7 ^ x; // <-- original
c99dc845 MHS	97	uint8_t z2 = ((r & ~r_ls2) >> 4) ^ (r_and_ls2 >> 3) ^ r ^ x;
	98
	99	return (z0 & 4) \| (z1 & 2) \| (z2 & 1);
	100	}
	101	*/
	102
17505ce2	103	void opt_successor(const uint8_t k, State s, bool y, State *successor) {
c99dc845 MHS	104	uint8_t Tt = 1 & opt_T(s);
	105
	106	successor->t = (s->t >> 1);
	107	successor->t \|= (Tt ^ (s->r >> 7 & 0x1) ^ (s->r >> 3 & 0x1)) << 15;
	108
	109	successor->b = s->b >> 1;
	110	successor->b \|= (opt_B(s) ^ (s->r & 0x1)) << 7;
	111
17505ce2	112	successor->r = (k[opt__select(Tt, y, s->r)] ^ successor->b) + s->l ;
17505ce2	113	successor->l = successor->r + s->r;
c99dc845 MHS	114	}
c99dc845 MHS	115
17505ce2	116	void opt_suc(const uint8_t k, State s, uint8_t *in, uint8_t length, bool add32Zeroes) {
c99dc845	117	State x2;
17505ce2	118	for (int i = 0; i < length; i++) {
17505ce2	119	uint8_t head;
c99dc845	120	head = 1 & (in[i] >> 7);
17505ce2	121	opt_successor(k, s, head, &x2);
c99dc845 MHS	122
c99dc845 MHS	123	head = 1 & (in[i] >> 6);
17505ce2	124	opt_successor(k, &x2, head, s);
c99dc845 MHS	125
c99dc845 MHS	126	head = 1 & (in[i] >> 5);
17505ce2	127	opt_successor(k, s, head, &x2);
c99dc845 MHS	128
c99dc845 MHS	129	head = 1 & (in[i] >> 4);
17505ce2	130	opt_successor(k, &x2, head, s);
c99dc845 MHS	131
c99dc845 MHS	132	head = 1 & (in[i] >> 3);
17505ce2	133	opt_successor(k, s, head, &x2);
c99dc845 MHS	134
c99dc845 MHS	135	head = 1 & (in[i] >> 2);
17505ce2	136	opt_successor(k, &x2, head, s);
c99dc845 MHS	137
c99dc845 MHS	138	head = 1 & (in[i] >> 1);
17505ce2	139	opt_successor(k, s, head, &x2);
c99dc845 MHS	140
c99dc845 MHS	141	head = 1 & in[i];
17505ce2	142	opt_successor(k, &x2, head, s);
c99dc845	143	}
61fe9073	144	//For tag MAC, an additional 32 zeroes
17505ce2	145	if (add32Zeroes) {
	146	for(int i = 0; i < 16; i++) {
	147	opt_successor(k, s, 0, &x2);
	148	opt_successor(k, &x2, 0, s);
61fe9073	149	}
17505ce2	150	}
c99dc845 MHS	151	}
c99dc845 MHS	152
17505ce2	153	void opt_output(const uint8_t k, State s, uint8_t *buffer) {
	154	State temp = {0, 0, 0, 0};
	155	for (uint8_t times = 0; times < 4; times++) {
	156	uint8_t bout = 0;
c99dc845	157	bout \|= (s->r & 0x4) << 5;
17505ce2	158	opt_successor(k, s, 0, &temp);
c99dc845	159	bout \|= (temp.r & 0x4) << 4;
17505ce2	160	opt_successor(k, &temp, 0, s);
c99dc845	161	bout \|= (s->r & 0x4) << 3;
17505ce2	162	opt_successor(k, s, 0, &temp);
c99dc845	163	bout \|= (temp.r & 0x4) << 2;
17505ce2	164	opt_successor(k, &temp, 0, s);
c99dc845	165	bout \|= (s->r & 0x4) << 1;
17505ce2	166	opt_successor(k, s, 0, &temp);
c99dc845	167	bout \|= (temp.r & 0x4) ;
17505ce2	168	opt_successor(k, &temp, 0, s);
c99dc845	169	bout \|= (s->r & 0x4) >> 1;
17505ce2	170	opt_successor(k, s, 0, &temp);
c99dc845	171	bout \|= (temp.r & 0x4) >> 2;
17505ce2	172	opt_successor(k, &temp, 0, s);
c99dc845 MHS	173	buffer[times] = bout;
c99dc845 MHS	174	}
c99dc845 MHS	175	}
c99dc845 MHS	176
17505ce2	177	void opt_MAC(uint8_t k, uint8_t input, uint8_t *out) {
c99dc845	178	State _init = {
17505ce2	179	((k[0] ^ 0x4c) + 0xEC) & 0xFF,// l
	180	((k[0] ^ 0x4c) + 0x21) & 0xFF,// r
	181	0x4c, // b
	182	0xE012 // t
	183	};
c99dc845	184
17505ce2	185	opt_suc(k, &_init, input, 12, false);
c99dc845	186	//printf("\noutp ");
17505ce2	187	opt_output(k, &_init, out);
c99dc845	188	}
17505ce2	189
c99dc845 MHS	190	uint8_t rev_byte(uint8_t b) {
	191	b = (b & 0xF0) >> 4 \| (b & 0x0F) << 4;
	192	b = (b & 0xCC) >> 2 \| (b & 0x33) << 2;
	193	b = (b & 0xAA) >> 1 \| (b & 0x55) << 1;
17505ce2	194	return b;
c99dc845	195	}
17505ce2	196
	197	void opt_reverse_arraybytecpy(uint8_t dest, uint8_t src, size_t len) {
	198	for (size_t i = 0; i < len; i++) {
c99dc845	199	dest[i] = rev_byte(src[i]);
17505ce2	200	}
c99dc845 MHS	201	}
c99dc845 MHS	202
17505ce2	203	void opt_doReaderMAC(uint8_t cc_nr_p, uint8_t div_key_p, uint8_t mac[4]) {
61fe9073	204	static uint8_t cc_nr[12];
17505ce2	205	opt_reverse_arraybytecpy(cc_nr, cc_nr_p, 12);
	206	uint8_t dest[] = {0, 0, 0, 0, 0, 0, 0, 0};
	207	opt_MAC(div_key_p, cc_nr, dest);
61fe9073	208	//The output MAC must also be reversed
17505ce2	209	opt_reverse_arraybytecpy(mac, dest, 4);
61fe9073 MHS	210	return;
61fe9073 MHS	211	}
17505ce2	212
17505ce2	213	void opt_doTagMAC(uint8_t cc_p, const uint8_t div_key_p, uint8_t mac[4]) {
61fe9073	214	static uint8_t cc_nr[8+4+4];
17505ce2	215	opt_reverse_arraybytecpy(cc_nr, cc_p, 12);
	216	State _init = {
	217	((div_key_p[0] ^ 0x4c) + 0xEC) & 0xFF,// l
	218	((div_key_p[0] ^ 0x4c) + 0x21) & 0xFF,// r
	219	0x4c, // b
	220	0xE012 // t
	221	};
	222	opt_suc(div_key_p, &_init,cc_nr, 12, true);
	223	uint8_t dest[] = {0, 0, 0, 0};
	224	opt_output(div_key_p, &_init, dest);
61fe9073	225	//The output MAC must also be reversed
17505ce2	226	opt_reverse_arraybytecpy(mac, dest, 4);
61fe9073	227	return;
61fe9073	228	}
17505ce2	229
61fe9073 MHS	230	/**
	231	* The tag MAC can be divided (both can, but no point in dividing the reader mac) into
	232	* two functions, since the first 8 bytes are known, we can pre-calculate the state
	233	* reached after feeding CC to the cipher.
	234	* @param cc_p
	235	* @param div_key_p
	236	* @return the cipher state
	237	*/
17505ce2	238	State opt_doTagMAC_1(uint8_t cc_p, const uint8_t div_key_p) {
61fe9073	239	static uint8_t cc_nr[8];
17505ce2	240	opt_reverse_arraybytecpy(cc_nr, cc_p, 8);
	241	State _init = {
	242	((div_key_p[0] ^ 0x4c) + 0xEC) & 0xFF,// l
	243	((div_key_p[0] ^ 0x4c) + 0x21) & 0xFF,// r
	244	0x4c, // b
	245	0xE012 // t
	246	};
	247	opt_suc(div_key_p, &_init, cc_nr, 8, false);
61fe9073 MHS	248	return _init;
61fe9073 MHS	249	}
17505ce2	250
61fe9073 MHS	251	/**
	252	* The second part of the tag MAC calculation, since the CC is already calculated into the state,
	253	* this function is fed only the NR, and internally feeds the remaining 32 0-bits to generate the tag
	254	* MAC response.
	255	* @param _init - precalculated cipher state
	256	* @param nr - the reader challenge
	257	* @param mac - where to store the MAC
	258	* @param div_key_p - the key to use
	259	*/
17505ce2	260	void opt_doTagMAC_2(State _init, uint8_t nr, uint8_t mac[4], const uint8_t div_key_p) {
17505ce2	261	static uint8_t _nr[4];
61fe9073	262	opt_reverse_arraybytecpy(_nr, nr, 4);
17505ce2	263	opt_suc(div_key_p, &_init, _nr, 4, true);
	264	//opt_suc(div_key_p, &_init,nr, 4, false);
	265	uint8_t dest[] = {0, 0, 0, 0};
	266	opt_output(div_key_p, &_init, dest);
c99dc845	267	//The output MAC must also be reversed
17505ce2	268	opt_reverse_arraybytecpy(mac, dest, 4);
c99dc845 MHS	269	return;
c99dc845 MHS	270	}